Big Data Startup Eyes Genome Analysis In 4 Hours

7 Big Data Solutions Try To Reshape Healthcare

A startup company has developed a scalable system that purportedly increases the speed of processing and analyzing human genomic sequences by a factor of 100, cutting analysis time down to four hours or less, and potentially making genomics and personalized medicine accessible outside of academia.

Bina Technologies, based in Redwood City, Calif. and founded in 2011 as a spin-out from Stanford University and the University of California at Berkeley, on Tuesday officially launched the Bina Genomic Analysis Platform. The system marries a scalable set of Intel-based servers with software that, according to senior VP of business development Mark Sutherland, "assembles a jigsaw puzzle with 1 billion pieces."

Bina is displaying the platform this week at the Advances in Genomic Biology & Technology conference in Marco Island, Fla.

The cost of sequencing a genome has fallen from $95.3 million in September 2001 to $6,618 as of October 2012, according to the U.S. government's National Human Genome Research Institute. A full sequence still cost more than $7 million as recently as late 2007. "In the last five years, we have seen a dramatic drop in the price of doing DNA sequencing," Sutherland told InformationWeek Healthcare.

[ IBM's cognitive computing technology could play a key role in solving healthcare's most urgent challenges. Read more: IBM's Watson Could Be Healthcare Game Changer. ]

But a complete genome sequence produces half a terabyte of data, and someone needs to mine and interpret all that information to make it suitable for specific research and treatment applications. "The data flow becomes very unwieldy," Sutherland explained.

"We operate at this intersection between big data and genomics," Sutherland said. "We see opportunity in bridging the gulf between technology providers and those that wish to consume and use the information." Sutherland explained that this gap historically has been bridged by open-source software in academia. "But in reality, this is not scalable or cost-effective."

Pilot customers for the Bina Genomic Analysis Platform have included the Stanford University School of Medicine's Department of Genetics and the Veterans Affairs Palo Alto (Calif.) Health Care System hospital. Prior to the Bina installation, Sutherland said, it typically took 10 days to go through one genome sample, in part because those institutions did not have a dedicated system and were using six different open-source software programs that were not linked.

"We are bringing the cost of analysis down to a few hundred dollars," Sutherland said. This, he believes, makes it more attractive to health insurance companies and to self-pay customers.

Next month, Bina expects to make an announcement related to care in neonatal intensive care units. Newborns who spend time in NICUs often are there because of "severe genetic distress," according to Sutherland, and genomic testing can help pinpoint ailments. "You have to know [what the problem is] within the first 48 to 72 hours for the physician to take action that can be successful," he said. "You can't wait two weeks for results."

Sutherland said genomic analysis will also be "super important" in the treatment of cancer, many forms of which have genetic bases. "To understand one person's genome well, you need to study 1,000 genomes or more [in order to find a genetic base for diseases]," explained Rupert Yip, Bina's director of project management. "This hasn't been possible before."

Users of the Bina Genomic Analysis Platform initially will be in "core labs" at academic medical centers, Sutherland said, but soon sequencing will begin to show up in community hospitals and even some independent laboratories. He predicted that the number of DNA sequencers in the U.S. will quadruple from about 2,000 today to 8,000 by 2015, with each one capable of decoding about one genome per day. That means that there will be 1 million genomes needing analysis by next year.

"We really see this making it into the clinic in the next 2 to 3 years," Yip said.

As large healthcare providers test the limits, many smaller groups question the value. Also in the new, all-digital Big Data Analytics issue of InformationWeek Healthcare: Ask these six questions about natural language processing before you buy. (Free with registration.)